Re: [PATCH net-next v4] net: ipv4: add support for ECMP hash policy choice

On 16/03/17 18:49, Nikolay Aleksandrov wrote:

On 16/03/17 18:41, Stephen Hemminger wrote:

quoted

On Thu, 16 Mar 2017 15:28:00 +0200
Nikolay Aleksandrov [off-list ref] wrote:

quoted

diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index d6880a6149ee..62c4f94923e5 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c

@@ -1004,6 +1004,15 @@ static struct ctl_table ipv4_net_table[] = {
 		.extra1		= &zero,
 		.extra2		= &one,
 	},
+	{
+		.procname	= "fib_multipath_hash_policy",
+		.data		= &init_net.ipv4.sysctl_fib_multipath_hash_policy,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &one,
+	

Rather than having magic integer values, it would be better to use
strings (like TCP congestion control).  Especially if you want to support
more values in the future.

With strings we'll need two sysctls - one to export the available ones, and one to set.
I too am not happy with plain integers as I've said in v1 of this patch but if people
are okay with having two new sysctls exported then I don't mind reworking it with strings.
Or any other ideas are welcome.

quoted

Also what about IPv6?

Patches are welcome, too. :-)
We can update it at any time, IPv4 and 6 have been different for a very long time
and it's not the point of this patch to bring them closer, though IPv6 already
does L4 by default.

Just to clarify, we're interested in getting IPv4 and 6 on the same level and will
work towards that but let's take it a step at a time.